This episode delves into the high-stakes clash between free trade principles and national security imperatives, analyzing how tariffs and export controls, particularly targeting China and AI technologies like Nvidia chips, create critical uncertainties for Crypto AI investors.
Navigating a World of Complex Systems and Policy Shifts
- Brad Gerstner opens by expressing concern over the sheer number of large-scale, simultaneous policy changes attempted by the current administration, spanning global trade, taxes, deficits, and geopolitical conflicts (Ukraine, Middle East).
- He frames this as a complex system challenge, highlighting the difficulty in predicting outcomes when multiple major variables are altered at once. Brad notes, “the probability that you're going to land this plane in just a really smooth landing I think is challenging to see.”
- Bill Gurley concurs, referencing his time at the Santa Fe Institute, which studies complex systems. He emphasizes the unpredictability arising from unknown interdependencies between variables, making accurate macroeconomic forecasting exceptionally difficult.
- Strategic Implication: The high degree of policy uncertainty across multiple fronts (trade, fiscal, geopolitical) increases systemic risk, demanding caution and adaptability from investors navigating tech and crypto markets.
The Tangible Impacts of Trade Tensions
- The discussion highlights immediate consequences, such as a potential $7 billion revenue hit to Meta from reduced advertising by Chinese merchants like Temu, illustrating the downstream effects of trade friction on even seemingly unrelated tech giants.
- Brad notes the market's anxiety, reflected in a 15% downturn since early April, coinciding with escalating trade rhetoric and actions.
Revisiting the Tariff Debate: Tactical vs. Structural
- The conversation revisits the administration's approach to tariffs, contrasting a "tactical, narrow view" (targeted tariffs for critical industries) with the "nuclear Navarro approach" (high, broad tariffs potentially replacing income tax). Peter Navarro is an economist known for advocating aggressive protectionist trade policies.
- Recent escalations include China's export ban on rare earths (April 5th) and the US retaliating with an export ban on Nvidia H20 chips. Rare earths are a group of elements critical for magnets and electronics. The Nvidia H20 is an AI accelerator chip specifically designed to comply with earlier US export restrictions to China.
- Brad points out the political tension, noting the historical Republican stance against tariffs and growing concern on Capitol Hill about the current trajectory's economic impact.
The Case Against Tariffs: Lessons from History
- Bill Gurley introduces a historical perspective, playing a clip of Ronald Reagan warning against protectionism. Reagan argued tariffs, like the Smoot-Hawley Tariff Act of 1930 which significantly raised US tariffs on imported goods, deepen economic downturns.
- Reagan's points, echoed by Bill:
- Tariffs initially seem patriotic but lead to domestic industries becoming reliant and less innovative.
- They inevitably trigger retaliation and trade wars.
- Reduced competition and artificially high prices shrink markets and destroy jobs.
- Bill expresses skepticism towards the current tariff strategy, citing alignment with established economic thought (e.g., Ricardo Hausman, Niall Ferguson) suggesting long-term negative consequences like inflation and reduced innovation. He finds the arguments for broad tariffs violate his priors learned over decades.
National Security vs. Free Trade: The Mearsheimer Argument
- Brad introduces the counterargument, channeling thinkers like John Mearsheimer (a prominent international relations scholar of the realist school) and potentially David Sachs. This view posits that in a great power struggle with a peer rival like China, national security must supersede free trade principles.
- This perspective justifies tariffs and controls to build resilience in critical sectors like chip fabs, rare earths, auto manufacturing, and pharmaceuticals.
- Crypto AI Relevance: This debate directly impacts the supply chain for computing hardware (GPUs, TPUs) essential for both AI model training and crypto mining/validating, influencing hardware costs and availability.
Critiquing the Current Approach: Consistency and Alternatives
- Bill argues that even if tactical resilience is the goal, the current "tone and bravado" feel inconsistent with a targeted approach. He stresses that resilience can involve allies (like Vietnam) and requires policy consistency for businesses to adapt.
- He references a Thomas Friedman op-ed suggesting the US may be falling behind in advanced manufacturing and might even need to learn from China via joint ventures (JVs).
- Brad agrees tactical re-onshoring of critical capabilities (chip fabs, pharma, rare earths) makes sense but criticizes the "nuclear style tariff war against every country," finding the sequencing confusing.
- He advocates shifting focus towards an "accelerationist agenda"—proactive measures like AI funds, tax incentives, and deregulation to boost domestic production, rather than just imposing tariffs.
Government Role: Picking Winners vs. Removing Friction
- Bill pushes back on government-led industrial policy, warning against "regulatory capture" (where regulatory agencies become dominated by the industries they regulate) and the government's poor track record at picking winners, especially given the massive private capital flowing into AI.
- He suggests government efforts are better focused on removing friction, citing the 4x higher cost of building nuclear fission reactors in the US compared to South Korea or China as an area needing reform.
- Brad concurs, emphasizing accelerating US capabilities by removing "needless regulation."
- Bill shares an anecdote (via Friedman/Ezra Klein) about China potentially fostering innovation by funding many competing startups (e.g., in solar) rather than picking a single winner, using market forces internally.
Export Controls: The Nvidia H20 and Rare Earth Bans
- The discussion sharpens focus on export controls – license requirements acting as trade embargoes on specific goods. Examples: China's rare earth ban (affecting magnets in EVs, phones, computers) and the US ban on Nvidia H20 chip sales to China.
- While China mines ~60% of rare earths, it refines ~90%, giving its ban significant leverage, described by some as a potential "kill shot" against US industrial production.
- Brad notes Trump's recent signals suggesting tariffs might be walked back, but highlights the distinct danger of export bans.
- Bill sees the retaliation as predictable, citing Reagan and Milton Friedman's warnings about the escalatory nature of trade restrictions. Friedman famously compared retaliatory tariffs to shooting more holes in your own boat.
The H20 Ban: Self-Defeating Strategy?
- Analysis suggests the US ban on Nvidia's H20 chip (already a "nerfed" chip designed for export compliance) may primarily benefit Huawei, China's domestic champion.
- Reports indicate Huawei's Ascend chips (like the Cluster Matrix 384) are already near-frontier, potentially closing the gap with the H20.
- The ban removes Nvidia/CUDA competition within China, potentially granting Huawei monopoly profits and accelerating its development as Chinese AI firms are forced onto its platform. CUDA (Compute Unified Device Architecture) is Nvidia's parallel computing platform and API model.
- Bill highlights the irony: “How good must it feel to be Nvidia... that you become a pawn in these games that you had nothing to do with creating.”
- Brad argues that while restricting cutting-edge chips (like the GB200) to China is defensible on national security grounds, banning the H20 seems counterproductive. It eliminated competition for Huawei, denied Nvidia profits (and US tax revenue), and may not significantly hinder China's AI progress given Huawei's capabilities and China's massive investments in compute and power infrastructure (e.g., nuclear).
- Strategic Implication: Export controls on AI hardware create significant volatility for chipmakers like Nvidia and downstream impacts for AI developers and crypto operations reliant on GPU availability and cost. The effectiveness and unintended consequences (like boosting competitors) must be carefully weighed.
The Flawed Logic of "Winning the AI War"
- Bill strongly critiques the prevalent rhetoric of needing to "win the AI war," arguing it stems from flawed "finite game" thinking applied to an "infinite game" like technological development. Finite games have clear winners, losers, and endpoints (like sports); infinite games continue indefinitely (like business or innovation).
- He asserts, “I don't know what that means [winning the AI war]. And if I guess as to what it means, I don't think it's possible.”
- He points out that even China hawks concede China's AI progress cannot be permanently stopped. Innovation spreads rapidly, citing OpenAI building on Google's work as a recent example.
- The belief that the US can somehow "increase the gap" or achieve a definitive win ignores the reality of global interdependence and China's own significant innovative capacity. Eric Schmidt and others note the AI gap may actually be shrinking.
- Crypto AI Relevance: Viewing AI development as a zero-sum "war" can lead to policies (like broad export bans or potential restrictions on open-source models like DeepSeek) that fragment the global ecosystem, potentially hindering access to cutting-edge research and tools for everyone, including the crypto AI space.
Potential Escalation and Alienating Allies
- The speakers warn that aggressive US policies risk alienating key allies and partners. Trying to extend controls to companies like ASML (Dutch) or TSMC (Taiwanese) could backfire.
- Pushing allies away while China adopts a more "statesmanlike" tone could inadvertently drive other nations closer to China, undermining US influence and the goal of having the world run on US-centric technology stacks.
- The potential ban on models from companies like DeepSeek (a Chinese AI research lab known for strong open-source models) is seen as another step down a slippery slope, potentially pushing regions like Europe and South America towards non-US AI ecosystems.
Market Volatility and Economic Uncertainty
- The discussion shifts to the struggling market (S&P -10%, Nasdaq -20% peak-to-trough), high volatility (VIX > 30), and concerns about Treasury demand and the dollar.
- Brad notes the dramatic sentiment shift since early year optimism (citing Stan Druckenmiller) to current widespread concern, reflected in significant drawdowns for major tech stocks (Apple, Nvidia, Google, Tesla).
- He observes that consensus S&P earnings expectations have barely fallen despite the chaos, suggesting potential downside risk if uncertainty persists and impacts corporate decision-making further. CEOs are reportedly cautious and delaying investments.
- Brad believes the market needs policy predictability soon. He anticipates an eventual landing zone with moderate tariffs (5-10% ROW, 25-30% China) and removal of export bans, but acknowledges significant headwinds if uncertainty drags on.
- Investor Takeaway: Current market conditions reflect deep uncertainty driven by trade policy. While dips may present opportunities, significant directional bets carry high risk until policy clarity emerges. Portfolio positioning should likely remain cautious (bottom third of normal risk exposure, as Brad suggests).
Conclusion: Focus on Acceleration, Not Obstruction
- The core strategic insight is that attempting to win an "AI war" by hindering China through broad tariffs and export controls is likely a counterproductive, infinite game strategy. It risks alienating allies, boosting competitors, and ultimately failing to stop China's progress.
- Crypto AI investors and researchers should monitor policy shifts closely. A pivot towards tactical re-onshoring and domestic acceleration, rather than broad protectionism, would signal greater market stability and a more favorable environment for tech innovation.
